Computer Graphics - Introduction - Philipp Slusallek Computer - - PowerPoint PPT Presentation
Computer Graphics - Introduction - Philipp Slusallek Computer Graphics WS 2019/20 Philipp Slusallek Overview Today Administrative stuff History of Computer Graphics (CG) Next lecture Overview of Ray Tracing Computer
Computer Graphics WS 2019/20 Philipp Slusallek
Computer Graphics WS 2019/20 Philipp Slusallek
– Administrative stuff – History of Computer Graphics (CG)
– Overview of Ray Tracing
Computer Graphics WS 2019/20 Philipp Slusallek
– Applied Computer Science (Praktische Informatik) – Lectures in English
– Mon 10:00-12:00h, HS 01, E1.3 – Thu 8:00-10:00h, HS 01, E1.3 (suggestion: 8:30-10:00h)
– 9 credit points
– http://graphics.cg.uni-saarland.de/courses/ – Schedule, slides as PDF, etc. – Literature, assignments, other information
– [Do not forget to sign-out in time before the exams, if you need to]
Computer Graphics WS 2019/20 Philipp Slusallek
– Philipp Slusallek
– Arsène Pérard-Gayot – E1.1, Room E13, Tel. 3837, Email: perard@cg.uni-saarland.de
– Julius Kilger (juliuskilger@posteo.de) – Joschua Loth (s8joloth@stud.uni-saarland.de) – Henrik Philippi (s8hephil@stud.uni-saarland.de)
Computer Graphics WS 2019/20 Philipp Slusallek
Computer Graphics WS 2019/20 Philipp Slusallek
– Theoretical & programming assignments – You will incrementally build your own ray tracing system – This will be the basis for the Rendering Competition
– Results of the exercises will contribute to the final grade – Bonus points (towards the exam) are possible
– Theoretical: In paper form (beginning of lecture) or PDF per email – Code: See exercise sheet or Web page (usually by email to tutor)
– Discuss lectures and any issues you might have with TAs
– Each one must be able to present and explain his/her results! – Please state who did what!!!
Computer Graphics WS 2019/20 Philipp Slusallek
– Counts 30% towards final grade (with +20% bonus points)
– Counts 10% towards final grade – Grading: Artistic quality (jury) – Groups of max. 2 students (but higher requirements then)
– Mid-term (exam prereq.), counts 20% towards final grade – Final exam counts 40% towards final grade – Minimum: 50% to pass (in each of the above)
– 0% of assignment grade on first attempt – Possibility to fail the entire course if repeated
– Oral exam (if possible) at the end of the semester break
Computer Graphics WS 2019/20 Philipp Slusallek
– Create a realistic image of a virtual environment – Incorporate additional technical features into your ray tracer – Bonus points count towards exam – Creative design of a realistic and/or aesthetic 3D scene – Modeling and shading
– You can work on it during the entire course – Deadline will be announced (see Web page)
– One rendered image – Web page with technical detail info
Computer Graphics WS 2019/20 Philipp Slusallek
Computer Graphics WS 2019/20 Philipp Slusallek
Computer Graphics WS 2019/20 Philipp Slusallek
– John Hughes, et al.: Computer Graphics – Principles and Practice, Addison-Wesley, 3. Ed, 2013 – Peter Shirley: Fundamentals in CG, 4. Ed, AK Peters, 2016 – Matt Pharr, Wenzel Jakob, Greg Humphreys: Physically Based Rendering : From Theory to Implementation, Morgan Kaufmann Series, 3. Ed., 2016, now freely available: http://www.pbr-book.org/
– Andrew Glassner: An Introduction to Ray-Tracing, Academic Press, 1989 – David Ebert: Texturing and Modeling – A procedural approach, Morgan Kaufmann, 2003 – T
Companion, Morgan Kaufmann, 2000
– Thomas Akenine-Möller, Eric Haines, Real-Time Rendering, AK Peters, 2nd Ed., 2002 – John M. Kessenich, et al., OpenGL Programming Guide, Addison- Wesley, 9. Ed., 2016
Computer Graphics WS 2019/20 Philipp Slusallek
Computer Graphics WS 2019/20 Philipp Slusallek
Geometric Modeling Rendering Simulation & Animation
Perception
Inverse Rendering
CAD/CAM/CAE
Computer Graphics WS 2019/20 Philipp Slusallek
Rendering Modeling Animation Visualization Imaging GUI VR/AR Digital Media Plotting Printer Color Management Computer Vision Computer Architecture Languages Systems Computer Games Compression Mathematical Modeling And, and, and, ....
Computer Graphics WS 2019/20 Philipp Slusallek
Computer Graphics WS 2019/20 Philipp Slusallek
Max-Planck Institutes University Business Units Blue-Sky Research Basic Research Applied Research Produkt Prototype Industry Research
Valley of DeathTM
Intel-VCI 1 Research 10 Engineering 100 Start-Ups (new IT-Incubator Saar) DFKI ASR Engineers Researchers Demonstrator
− „Computer with Eyes, Ears, and Common Sense“
− Largest AI research center worldwide (founded in 1988) − Germany’s leading research center for innovative SW technologies − 6 sites in Germany
− 18 research areas, 10 competence centers, 7 living labs − More than 575 core research staff (>1050 total) − Revenues of ~50 M€ (2018) − More than 90 spin-offs
DFKI: The World´s Largest Center for Research & Application in AI
Saarbrücken Berlin Bremen Osnabrück Kaiserslautern Oldenburg
Deutschland GmbH
‘Blue Sky‘ Basic Research Commercialization/ Exploitation
Labs at the University
Application- inspired Basic Research Applied Research and Development Transfer Projects DFKI projects for external clients and shareholders Spin-off Companies with DFKI equity DFKI projects for federal government, EU DFKI projects for state governments, clients and shareholders External Clients Shareholders
The verticalspecialisationof DFKI
Max Planck Society Fraunhofer Helmholtz Society
The entire innovation chain in the horizontal spectrum of DFKI
Deep knowledge and excellence in
Broad Methodological and Systems Competence in Artificial Intelligence
Deep Domain Knowledge in an Area of Application
DFKI Employees
Application-Oriented Basic Research
Applied R&D and Transfer
Large T est- and Demonstration Centers
Deep Scientific Expertise in AI Technology
9 DFKI Heads of Research Labs 10 Associated and Supernumerary Professors 10 Heads of Research Groups / Living Lab Leaders at DFKI
Prof. Rolf Drechsler Prof. Philipp Slusallek Prof. Andreas Dengel Prof. Didier Stricker Prof. Paul Lukowicz Prof. Frank Kirchner Prof. Antonio Krüger Prof. Josef van Genabith Prof. Tim E. Güneysu Prof. Hans Uszkoreit Prof. Udo Frese Prof. Dieter Hutter Prof. Christoph Lüth Prof. Jana Koehler Prof. Martin Ruskowski Prof. Peter Loos Prof. Volker Markl Prof. Wolfgang Maaß Prof. Gesche Joost Prof. Joachim Hertzberg Prof. Sebastian Möller Prof. Oliver Zielinski Prof. Oliver Thomas Prof. Hans D. Schotten Prof. Klaus-Dieter Althoff Prof. Peter Fettke Prof. Stephan Busemann Prof. Günter Neumann Prof. Wolfgang Wahlster
Plan-Based Robot Control Robotics Innovation Center Institute for Information Systems Smart Service Engineering Intelligent Analytics for Massive Data Intelligent Networks Multilinguality and Language Technology
European Media Lab
Bundesamt f ür Sicherheit in der Inf ormationstechnik
BSI
Competence Center Informatik
BASF Aktiengesellschaft
Verlagsgruppe Georg v on Holzbrinck
Saarländische Polizei
Bundesministerium f ür Wirtschaf t u. Arbeit
5 4 2 1 2 19 1 1 24 7 1 23 1 9 46 1 5 1 2 4 2 1 4 4 3 2 1 4 1 1 6 1 5 35 5 1 1 2 1 10 1 3 1 7 3 2 2 1 4 2 1 1 5 2 3 1 1 3 3 2 2
AI, Graphics/Simulation, High-Performance Computing
Important German, European & International Cooperations:
Research Teams
Distributed & Web- Based Systems René Schubotz High-Performance Graphics & Computing Richard Membarth Computational 3D Imaging Tim Dahmen
Scientific Director
Intelligent Information Systems Matthias Klusch Multi-Agent Systems Klaus Fischer Smart Living Hilko Hoffmann
Application Domains
Autonomous Driving Christian Müller Industrie 4.0 Ingo Zinnikus Smart Living Hilko Hoffmann Autonomous Driving Christian Müller High-Performance Computing Richard Membarth Computational Sciences Tim Dahmen
AI Platform ML / DL
Strategy Board Christian Müller (Deputy) Silke Balzert-Walter (Consulting) Philipp Slusallek
Application- Driven Teams Hybrid & Symbolic AI
Philipp Slusallek, Christian Müller Philipp Slusallek, Christian Müller Philipp Slusallek (Co-Initiator), Silke Balzert-Walter Philipp Slusallek Hilko Hoffmann
Computer Graphics WS 2019/20 Philipp Slusallek
– Simulated/Digital Reality (graphics, interaction, simulation) – Multi-agent Systems (AI: perception, learning, reasoning, planning) – HPC (compiler, parallel/vector computing: CPU/GPU/FPGA) – Visualization Center (presentation, teaching/training, consulting)
– >40 PhDs and researchers (plus many HiWis, BS, MS) – Many publicly funded projects
, Metacca, ProThOS, HP-DLF, SmartMaaS, …
– Researcher and engineer positions
– Extremely broad industry network (Contacts & Jobs, etc.)
Computer Graphics WS 2019/20 Philipp Slusallek
Flexible Production Control Using Multiagent Systems Verification and Secure Systems (BSI-certified Evaluation Center) Physically-Based Image Synthese Scientific Visualisation GIS and Geo Visualization Reconstruction of Cultural Heritage Future City Planning and Management Large 3D Models and Environments Large Visualization Systems Intelligent Human Simulation in Production Web-based 3D Application (XML3D) Distributed Visualization on the Internet
Key product offered now by all major HW vendors: e.g. Intel (Embree), Nvidia (OptiX), AMD (Radeon Rays) , …
DFKI Agenten und Simulierte Realität 29
VCM now part of most commercial renders: e.g. RenderMan, V-Ray, Corona, …
Numerous patents and spin-off companies from our group: e.g. inTrace, Motama, xaitment, PXIO, …
Real-Time Ray Tracing Hardware is integral part of every Top Nvidia GPU starting end of 2018
Three Siggraph papers in 2019 alone!
Developer
Computer Vision DSL
AnyDSL Compiler Framework (Thorin)
Physics DSL … Ray Tracing DSL
Various Backends (via LLVM)
Parallel Runtime DSL
Impala Language & Unified Program Representation Layered DSLs CPUs GPUs FPGAs Accels
Saarstahl, Völklingen
DFKI multi-agent technology is running the steelworks, 24/7 for >12 years, 5 researchers transferred
DFKI Agenten und Simulierte Realität 41
(using autonomous driving as an example)
− Geometry, Appearance, Motion, Weather, Environment, …
− Especially in Critical Situations − Increasingly making use of (deep) machine learning
− Often little (good) data even for “normal” situations − Critical situations rarely happen in reality – per definition! − Extremely high-dimensional models
− Continuous benchmarking & validation (“Virtual Crash-Test“)
− E.g. driving millions of miles to gather data − Difficult, costly, and non-scalable
Reality Car
− Arbitrarily scalable (given the right platform) − But: Where to get the models and the training data from?
Reality Digital Reality Car Car
Teil-Modelle Teil-Modelle Teil-Modelle Reality Digital Reality Partial Models (Rules)
Modeling & Learning
Car Car
Model Learning
Geometry Material Behavior Motion Environment …
Szenarien Szenarien Teil-Modelle Teil-Modelle Teil-Modelle Reality Digital Reality Partial Models (Rules) Relevant Scenarios Concrete Instances of Scenarios
Configuration & Learning Modeling & Learning Coverage of Variability via Directed Search
Car Car
Model Learning Reasoning
Szenarien Szenarien Teil-Modelle Teil-Modelle Teil-Modelle Reality Digital Reality Partial Models (Rules) Simulation/ Rendering Relevant Scenarios Concrete Instances of Scenarios
Configuration & Learning Modeling & Learning Synthetic Sensor Data, Labels, … Adaptation to the Simulated Environment (e.g. used sensors)
Car Car
Model Learning Simulation & Learning Reasoning
Coverage of Variability via Directed Search
Szenarien Szenarien Teil-Modelle Teil-Modelle Teil-Modelle Reality Digital Reality Partial Models (Rules) Simulation/ Rendering Relevant Scenarios Concrete Instances of Scenarios
Configuration & Learning Modeling & Learning Synthetic Sensor Data, Labels, … Adaptation to the Simulated Environment (e.g. used sensors)
Car Continuous Validation & Adaptation Car
Model Learning Reasoning Validation / Adaptation / Certification
Simulation & Learning
Coverage of Variability via Directed Search
Coverage of Variability via Directed Search
Szenarien Szenarien Teil-Modelle Teil-Modelle Teil-Modelle Reality Digital Reality Partial Models (Rules) Simulation/ Rendering Relevant Scenarios Concrete Instances of Scenarios
Configuration & Learning Modeling & Learning Synthetic Sensor Data, Labels, … Adaptation to the Simulated Environment (e.g. used sensors)
Car Continuous Validation & Adaptation Car
Model Learning Reasoning Validation / Adaptation / Certification
Continuous Learning Loop Not just for Automated Driving: Works for any AI System where we can model its interaction with the environment Simulation & Learning
− E.g. Gunnar Johansson's Point Light Walkers (1974) − Significant interdisciplinary research (e.g. psychology)
− E.g. gender, age, weight, mood, ... − Based on minimal information
− Detect if pedestrian will cross the street − Parameterized motion model & style transfer − Predictive models & physical limits
− Clear motion differences when crossing the street
Crossing Bus
− Longer wavelength: Geometric optics (rays) not sufficient − Need for some wave optics
− Highly different goals
− Identifying “useful” specular paths (using VCM) − Guides samples to visible specular effects (e.g. indirect radar echos)
− Gaming: Very nice images (at 60+ Hz)
− Film & Marketing: Even nicer images (at hours per image)
− Both are being used for simulations for Autonomous Driving
− Lidar, radar, multi-spectral, polarization, measured materials, … − Need for “error bar per pixel” & validation − Existing engines unlikely to adapt to these fundamental changes
− Focused on physical accuracy & high throughput − Based on latest graphics research results (and GPU-HW)
Computer Graphics WS 2019/20 Philipp Slusallek
– Rendering, Modeling, Visualization, Animation, Imaging, …
– “Everything is possible” mentality – Progress driven by research & technology – Flexible transfer between research and industry
– Intel, Nvidia, AMD, Imagination, ARM, … – Automotive, aerospace, engineering, … – Entertainment: games, film, TV, animations, ...
– Digital Reality, Visualization, Industrie-4.0, Big Data, Smart Cities, …
– Relations to mathematics, physics, engineering, psychology, art, entertainment, …